Unsteady forces on circular cylinders in a cross-flow

A three-axis piezoelectric load cell was used to measure the local unsteady forces induced on cylinders placed in a cross-flow. In conjunction with this, a single hot-wire was used to traverse the wake at a fixed distance behind the cylinder so that correlations between the induced forces on the cylinder and the wake velocity could be calculated to provide insight into the character of the flow-induced unsteady forces. Experiments were carried out on both two-dimensional and finite-span cylinders at a Reynolds number of 46,000. For the two-dimensional cylinder case, substantial evidence was obtained to demonstrate that the strength of the vortex roll-up along the span was quite uniform. Consequently, the lift-velocity correlation along the span remained unchanged. On the other hand, there was a total lack of correlation between the fluctuating drag and the wake velocity, thus indicating that the drag signal was not quite periodic. In the finite-span cylinder case, the separated flow from the top edge of the cylinder was found to suppress vortex shedding along the span of the cylinder, destroyed its coherence and caused the wake flow to oscillate in the stream direction. This oscillation induced a significant fluctuating drag on the cylinder. Consequently, the fluctuating drag far exceeded the fluctuating lift and the wake velocity was found to correlate well with the drag and not with the lift. This correlation remained intact along the span of the cylinder. Finally, the rms fluctuating lift and drag forces were found to vary along the cylinder span, with the lift increasing and the drag decreasing as the base of the cylinder is approached; thus suggesting that a submerged two-dimensional region exists near the base of the cylinder.List of symbols a span of active element on cylinder - C _D local rms drag coefficient, - C _L local rms lift coefficient, - C _D local mean drag coefficient - (C _D ) _2D spanwise-averaged mean drag coefficient for two dimensional cylinder - d diameter of cylinder (= 10.2 cm) - D fluctuating component of instantaneous drag - D local rms of fluctuating drag - E _D power spectrum of fluctuating drag, defined as - E _L power spectrum of fluctuating lift, defined as - E _U power spectrum of fluctuating streamwise velocity, defined as - f _L dominant frequency of lift spectrum - f _D dominant frequency of drag spectrum - f _u dominant frequency of velocity spectrum - h span of cylinder - H height of test section (= 30.5 cm) - L fluctuating component of instantaneous lift - L local rms of fluctuating lift - R _Du () cross-correlation function of streamwise velocity and local drag - R _Lu () cross-correlation function of streamwise velocity and local lift - Re Reynolds number, - S _L Strouhal number based on f _L , - S _D Strouhal number based on f _D , - S _U Strouhal number based on f _u , - t time - u fluctuating component of instantaneous streamwise velocity - u rms of streamwise fluctuating velocity - u rms of streamwise fluctuating velocity upstream of cylinder - U mean streamwise velocity - U mean stream velocity upstream of cylinder - x streamwise distance measured from axis of cylinder - y transverse distance measured from axis of cylinder - z spanwise distance measured from floor of test section - v kinematic viscosity of air - density of air - time lag in cross-correlation function - _D normalized spectrum of fluctuating drag - _L normalized spectrum of fluctuating lift - _U normalized spectrum of fluctuating streamwise velocity     

Two circular cylinders in cross-flow: A review

Pairs of circular cylinders immersed in a steady cross-flow are encountered in many engineering applications. The cylinders may be arranged in tandem, side-by-side, or staggered configurations. Wake and proximity interference effects, which are determined primarily by the longitudinal and transverse spacing between the cylinders, and also by the Reynolds number, have a strong influence on the flow patterns, aerodynamic forces, vortex shedding, and other parameters. This paper reviews the current understanding of the flow around two “infinite” circular cylinders of equal diameter immersed in a steady cross-flow, with a focus on the near-wake flow patterns, Reynolds number effects, intermediate wake structure and behaviour, and the general trends in the measurements of the aerodynamic force coefficients and Strouhal numbers. A primary focus is on the key experimental and numerical studies that have appeared since the last major review of this subject more than 20 years ago.     

Three-dimensional computation for flow-induced vibrations of an upstream circular cylinder in two tandem circular cylinders

It is well known from a lot of experimental data that fluid forces acting on two tandem circular cylinders are quite different from those acting on a single circular cylinder. Therefore, we first present numerical results for fluid forces acting on two tandem circular cylinders, which are mounted at various spacings in a smooth flow, and second we present numerical results for flow-induced vibrations of the upstream circular cylinder in the tandem arrangement. The two circular cylinders are arranged at close spacing in a flow field. The upstream circular cylinder is elastically placed by damper-spring systems and moves in both the in-line and cross-flow directions. In such models, each circular cylinder is assumed as a rigid body. On the other hand, we do not introduce a turbulent model such as the Large Eddy Simulation (LES) or Reynolds Averaged Navier-Stokes (RANS) models into the numerical scheme to compute the fluid flow. Our numerical procedure to capture the flow-induced vibration phenomena of the upstream circular cylinder is treated as a fluid-structure interaction problem in which the ideas of weak coupling is taken into consideration.     

Force coefficients and Strouhal numbers of three circular cylinders subjected to a cross-flow

In this paper, wind tunnel experiments were conducted to measure the mean force coefficients and Strouhal numbers for three circular cylinders of equal diameters in an equilateral-triangular arrangement when subjected to a cross-flow. These experiments were carried out at five subcritical Reynolds numbers ranging from 1.26 × 10⁴ to 6.08 × 10⁴. The pressure distributions on the surface of the cylinders were measured using pressure transducers. Furthermore, the hot-wire anemometer was employed to measure the vortex shedding frequencies behind each cylinder. Six spacing ratios (l/d) varying from 1.5 to 4 were investigated. It is observed that for l/d > 2, the upstream cylinder experiences a lower mean drag coefficient compared with the downstream cylinders. The minimum values of the drag coefficient for the downstream cylinders occur at l/d = 1.5 and l/d = 2, because there is no vortex shedding from the foregoing cylinders. Also, the value of the pressure coefficient behind the upstream cylinder reduces by increasing l/d. Moreover, by decreasing the value of l/d, the Strouhal number for the upstream cylinder increases. It can be concluded that the flow pattern and aerodynamic coefficients are basically dependent on l/d; in other words, decreasing l/d results in an increase in the effects of the flow interference between the cylinders.     

Characteristics and suppression of flow-induced vibrations of two side-by-side circular cylinders

A free-vibration experiment was conducted to examine flow-induced vibration (FIV) characteristics of two identical circular cylinders in side-by-side arrangements at spacing ratio T^⁎ (=T/D)=0.1–3.2, covering all possible flow regimes, where T is the gap spacing between the cylinders and D is the cylinder diameter. Each of the cylinders was two-dimensional, spring mounted, and allowed to vibrate independently in the cross-flow direction. Furthermore, an attempt to suppress flow-induced vibrations was undertaken by attaching flexible sheets at the rear stagnation lines of the cylinders. Based on the vibration responses of the two cylinders, four vibration patterns I, II, III and IV are identified at 0.1≤T^⁎<0.2, 0.2≤T^⁎≤0.9, 0.9<T^⁎<2.1 and 2.1≤T^⁎≤3.2, respectively. Pattern I is characterized by the two cylinders vibrating inphase^, with the maximum amplitudes occurring at the same reduced velocity Ur=10.47 almost two times that (Ur=5.25) for an isolated cylinder. Pattern II features no vibration generated for either cylinder. Pattern III exemplifies the occurrence of the maximum vibration amplitude of a cylinder at a smaller Ur and that of the other cylinder at a higher Ur compared to its counterpart in an isolated cylinder. Pattern IV represents each cylinder response resembling an isolated cylinder response; the vibrations of the two cylinders are, however, coupled inphase or antiphase. Linking maximum vibration amplitudes to fluctuating lift forces acting on fixed cylinders reveals that fluid–structure interactions between two fixed cylinders and between two elastic cylinders are not the same, even though vibration is not significant. As such, while two fixed cylinders generate narrow and wide wakes at 0.2≤T^⁎<1.7, two elastic cylinders do the same for a longer range of T^⁎ (0.2≤T^⁎<2.1). The flexible sheets effectively suppress FIV of the two cylinders in patterns III and IV, and reduce the vibration amplitude in pattern I. For the effectively controlled cases (patterns III and IV), the flexible sheet of each cylinder folds into a semicircle at the base, forming two free edges.     

Experiment on smooth,circular cylinders in cross-flow in the critical Reynolds number regime

Experiments were conducted for 2D circular cylinders at Reynolds numbers in the range of 1.73 × 10⁵–5.86 × 10⁵. In the experiment, two circular cylinder models made of acrylic and stainless steel, respectively, were employed, which have similar dimensions but different surface roughness. Particular attention was paid to the unsteady flow behaviors inferred by the signals obtained from the pressure taps on the cylinder models and by a hot-wire probe in the near-wake region. At Reynolds numbers pertaining to the initial transition from the subcritical to the critical regimes, pronounced pressure fluctuations were measured on the surfaces of both cylinder models, which were attributed to the excursion of unsteady flow separation over a large circumferential region. At the Reynolds numbers almost reaching the one-bubble state, it was noted that the development of separation bubble might switch from one side to the other with time. Wavelet analysis of the pressure signals measured simultaneously at θ = ±90° further revealed that when no separation bubble was developed, the instantaneous vortex-shedding frequencies could be clearly resolved, about 0.2, in terms of the Strouhal number. The results of oil-film flow visualization on the stainless steel cylinder of the one-bubble and two-bubble states showed that the flow reattachment region downstream of a separation bubble appeared not uniform along the span of the model. Thus, the three dimensionality was quite evident.     

Measurements of transverse forces on circular cylinders undergoing vortex-induced vibration

Experiments have been carried out on a circular cylinder, with and without helical strakes, free to respond in a direction transverse to a water flow. The Reynolds number range was between 3×10³ and 2.1×10⁴, the mass ratio was just above 0.8 and the fraction of critical damping was approximately 2×10⁻⁴. Measurements are presented of the response, the transverse fluid force and the phase angle between the response and the force, all as a function of reduced velocity. The straked cylinder is observed to respond over a narrow range of reduced velocity and its maximum amplitude is decreased by just over 60%, compared with a plain cylinder. The familiar phase jump that occurs for a plain cylinder did not occur with the straked one, with the phase close to zero over the entire reduced velocity range where response to vortex shedding occurred.     

Experimental investigation of flow-induced vibration on isolated and tandem circular cylinders fitted with strakes

The effect of varying the geometric parameters of helical strakes on vortex-induced vibration (VIV) is investigated in this paper. The degree of oscillation attenuation or even suppression is analysed for isolated circular cylinder cases. How a cylinder fitted with strakes behaves when immersed in the wake of another cylinder in tandem arrangement is also investigated and these results are compared to those with a single straked cylinder. The experimental tests are conducted at a circulating water channel facility and the cylindrical models are mounted on a low-damping air bearing elastic base with one degree-of-freedom, restricted to oscillate in the transverse direction to the channel flow. Three strake pitches (p) and heights (h) are tested: p=5, 10, 15d, and h=0.1, 0.2, 0.25d. The mass ratio is 1.8 for all models. The Reynolds number range is from 1000 to 10 000, and the reduced velocity varies up to 21. The cases with h=0.1d strakes reduce the amplitude response when compared to the isolated plain cylinder, however the oscillation still persists. On the other hand, the cases with h=0.2, 0.25d strakes almost completely suppress VIV. Spanwise vorticity fields, obtained through stereoscopic digital particle image velocimetry (SDPIV), show an alternating vortex wake for the p=10d and h=0.1d straked cylinder. The p=10d and h=0.2d cylinder wake has separated shear layers with constant width and no roll-up close to the body. The strakes do not increase the magnitude of the out-of-plane velocity compared to the isolated plain cylinder. However, they deflect the flow in the out-of-plane direction in a controlled way, which can prevent the vortex shedding correlation along the span. In order to investigate the wake interference effect on the strake efficiency, an experimental arrangement with two cylinders in tandem is employed. The centre-to-centre distance for the tandem arrangement varies from 2 to 6. When the downstream p=10d and h=0.2d cylinder is immersed in the wake of an upstream fixed plain cylinder, it loses its effectiveness compared with the isolated case. Although the oscillations have significant amplitude, they are limited, which is a different behaviour from that of a tandem configuration with two plain cylinders. For this particular case, the amplitude response monotonically increases for all gaps, except one, a trait usually found in galloping-like oscillations. SDPIV results for the tandem arrangements show alternating vortex shedding and oscillatory wake.     

Effect of acoustic resonance on the dynamic lift forces acting on two tandem cylinders in cross-flow

Direct measurements of the dynamic lift force acting on two tandem cylinders in cross-flow are performed in the presence and absence of acoustic resonance. The dynamic lift force is measured because it represents the integrated effect of the unsteady wake and therefore it is directly related to the dipole sound source generated by vortex shedding from the cylinder. Three spacing ratios inside the proximity interference region, L/D=1.75, 2.5 and 3 are considered. During the tests, the first transverse acoustic mode of the duct housing the cylinders is self-excited. In the absence of acoustic resonance, the measured dynamic lift coefficients agree with those reported in the literature. When the acoustic resonance is initiated, a drastic increase in the dynamic lift coefficient is observed, especially for the downstream cylinder. This can be associated with abrupt changes in the phase between the lift forces and the acoustic pressure. The dynamic lift forces on both cylinders are also decomposed into in-phase and out-of-phase components, with respect to the resonant sound pressure. The lift force components for the downstream cylinder are found to be dominant. Moreover, the out-of-phase component of the lift force on the downstream cylinder is found to become negative over two different ranges of flow velocity and to virtually vanish between these two ranges. Acoustic resonance of the first mode is therefore excited over two ranges of flow velocity separated by a non-resonant range near the velocity of frequency coincidence. It is therefore concluded that the occurrence of acoustic resonance is controlled by the out-of-phase lift component of the downstream cylinder, whereas the effect of the in-phase lift component is confined to causing small changes in the acoustic resonance frequency.     

Response and wake patterns of two side-by-side elastically supported circular cylinders in uniform laminar cross-flow

Vortex-induced vibrations (VIV) of two side-by-side elastically supported circular cylinders in a uniform flow with the Reynolds number of 100 are numerically investigated by using the immersed boundary method. The cylinders are constrained to oscillate in the cross-flow direction with a center-to-center spacing ratio $T/D$ ranging from 2 to 5. The structural damping is set to zero to enable large vibration amplitudes in the range of reduced velocity $U_r=3-10$. It is found that the proximity of the cylinders does not have a significant impact to the lock-in region and cylinder responses, except at a small spacing ratio of $T/D=2$. The critical spacing ratio is determined as $T/D=4$ and beyond that the interaction between the cylinders is negligible. The following six near-wake patterns are observed; the irregular pattern, in-phase flip-flopping pattern, out-of-phase flip-flopping pattern, in-phase-synchronized pattern, antiphase-synchronized pattern and the biased antiphase-synchronized pattern. These patterns are plotted in a plane of $U_r$ and $T/D$, together with approximate borderlines to distinguish one region from the others. The time histories, spectral features and wavelet transform contours of drag and lift forces are presented to elucidate the mechanisms of the in-phase and out-of-phase flip-flopping phenomena. It is established that the in-phase flip-flopping stems from the long-short near-wake pattern and its low-frequency flip-over, whereas the out-of-phase pattern originates from the large vortex shedding from the fictitious bluff-body with an augmented characteristic length.     

Flow/acoustic interactions of two cylinders in cross-flow

The noise generated by two tandem cylinders in a cross-flow (i.e., with the second in the wake of the first) has been investigated. Measurements of turbulence and of fluctuating pressure have been obtained between the two cylinders for different flow velocities and incident levels of turbulence. Although, for a number of cases, up to four peaks related to vortex shedding were evident in the spectrum, most measurements exhibited two peaks, a dominant one at the vortex-shedding frequency, with a secondary peak at twice this value. The measurements show that vortex generated noise is strongest at the mid-point between the cylinders and at the rear cylinder with levels of 130 dB. The harmonic component was strongest at the downstream cylinder where peak values of 110 dB were obtained. The nonlinear flow/acoustic interactions are examined using bispectral analysis to identify the quadratic interactions in the parameters. A novel quadratic modelling method is proposed and shown to be capable of both identifying and quantifying the nonlinear interactions which give rise to noise at harmonics of the vortex-shedding frequency.     

Flow-excited acoustic resonance of two side-by-side cylinders in cross-flow

The aeroacoustic response of two side-by-side circular cylinders in cross-flow is investigated experimentally. In order to investigate the effect of the gap between the cylinders on the acoustic resonance mechanism, six spacing ratios between the cylinders, in the range of T/D=1.25–3, have been investigated, where D is the diameter of the cylinders and T the centre-to-centre distance between them. Special attention is given to the intermediate spacing ratio range, which exhibits bistable flow regimes in the absence of resonance. During the tests, the acoustic cross-modes of the duct housing the cylinders are self-excited. For the intermediate spacing ratios, T/D=1.25, 1.35, 1.46 and 1.75, two distinct vortex-shedding frequencies at the off-resonance conditions are observed. These are associated with the wide and narrow wakes of the cylinders, as described in the literature. In this case, acoustic resonances occur at a Strouhal number, which is between those observed before the onset of resonance. The acoustic resonance synchronizes vortex shedding in the two wakes and thereby eliminates the bistable flow phenomenon. For large spacing ratios, T/D=2.5 and 3, vortex shedding occurs at a single Strouhal number at which the acoustic resonance is excited.     

Aspect ratio effect on natural convection in water near its density maximum temperature

The effect of the aspect ratio on natural convection in water subjected to density inversion has been investigated in this study. Numerical simulations of the two-dimensional, steady state, incompressible flow in a rectangular enclosure with a variety of aspect ratios, ranging from 0.125 to 100, have been accomplished using a finite element model. Computations cover Rayleigh numbers from 10³ to 10⁶. Results reveal that the aspect ratio, A, the Rayleigh number, Ra, and the density distribution parameter, R, are the key parameters to determine the heat transfer and fluid flow characteristics for density inversion fluids in an enclosure. A new correlation for predicting the maximum mean Nusselt number is proposed in the form of , with the constants a and b depending on density distribution number R. It is demonstrated that the aspect ratio has a strong impact on flow patterns and temperature distributions in rectangular enclosures. The stream function ratio Ψ_inv/|Ψ_reg| is introduced to describe quantitatively the interaction between inversional and regular convection. For R=0.33, the density inversion enhancement is observed in the regime near A=3.     

Experiments on the flow past long circular cylinders in a shear flow

This paper describes an experimental investigation of the flow past circular cylinders, with the mean flow perpendicular to the cylinder axis, at conditions that yield a strong three-dimensional behaviour. The experiments were carried out in the subcritical regime. Long cylinders with end plates were subjected to shear flow with a linear velocity profile in the spanwise direction generated by means of a curved gauze. It was concluded that spanwise cellular structures of vortex shedding emerged in the wake, more clearly for some boundary conditions than others. These structures are characterised by a portion of spanwise length, a cell, having constant shedding frequency over a time average, which implies that there were no vortex dislocations inside that cell during that time. These features were studied using flow visualisation and hot-film anemometry. Spectra of the local shedding frequency are shown, revealing the effect of the shear parameter (=0.02 and 0.04) and aspect ratio L/D (=20.6 and 8) on the stability and geometry of the cells at several Reynolds numbers in the range of 3.13×10³Re_m1.25×10⁴.     

A modified wake oscillator model for vortex-induced vibration of circular cylinders for a wide range of mass-damping ratio

In this paper, the behavior of an elastically mounted cylinder, subjected to vortex-induced vibrations (VIV), is investigated by a low-dimensional model. The classical wake oscillator model, as a standard model, predicts the behavior of the system at high mass-damping ratios but fails in modeling the system at low mass-damping ratios. A modified wake oscillator model is introduced in order to describe the response of the system over a wide range of mass-damping ratios. The results of this new model are compared to experimental results from the literature and shown to be in good agreement. The new model can describe most of the features of vortex-induced vibration phenomenology, such as the Griffin plot and lock-in domains.     

Fluid excitation of an oscillating circular cylinder in cross-flow

This paper presents an experimental technique developed to identify the fluid excitation force of a circular cylinder during Vortex-Induced Vibrations (VIV). To this end, an actuator is used to provide controlled damping and a Van der Pol model is proposed to describe the fluid excitation force, where the associated parameters are identified from experiments by a least square fitting. The Reynolds number was in the range 1650–4950, and the technique is validated with experimental free vibration data of an elastically supported circular cylinder. Also, comparison with free vibration measurements from other experiments carried out with similar Reynolds numbers is presented.